kvm.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Zenghui Yu <yuzenghui@huawei.com>
To: Marc Zyngier <maz@kernel.org>,
	<linux-arm-kernel@lists.infradead.org>,
	<kvmarm@lists.cs.columbia.edu>, <kvm@vger.kernel.org>,
	<linux-kernel@vger.kernel.org>
Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
	Jason Cooper <jason@lakedaemon.net>,
	Robert Richter <rrichter@marvell.com>,
	"Thomas Gleixner" <tglx@linutronix.de>,
	Eric Auger <eric.auger@redhat.com>,
	"James Morse" <james.morse@arm.com>,
	Julien Thierry <julien.thierry.kdev@gmail.com>,
	Suzuki K Poulose <suzuki.poulose@arm.com>
Subject: Re: [PATCH v5 16/23] irqchip/gic-v4.1: Eagerly vmap vPEs
Date: Tue, 17 Mar 2020 10:49:18 +0800	[thread overview]
Message-ID: <2817cb89-4cc2-549f-6e40-91d941aa8a5f@huawei.com> (raw)
In-Reply-To: <20200304203330.4967-17-maz@kernel.org>

Hi Marc,

On 2020/3/5 4:33, Marc Zyngier wrote:
> Now that we have HW-accelerated SGIs being delivered to VPEs, it
> becomes required to map the VPEs on all ITSs instead of relying
> on the lazy approach that we would use when using the ITS-list
> mechanism.
> 
> Signed-off-by: Marc Zyngier <maz@kernel.org>

Before GICv4.1, we use vlpi_count to evaluate whether the vPE has been
mapped on the specified ITS, and use this refcount to only issue VMOVP
to those involved ITSes. It's broken after this patch.

We may need to re-evaluate "whether the vPE is mapped" now that we're at
GICv4.1, otherwise *no* VMOVP will be issued on the v4.1 capable machine
(I mean those without single VMOVP support).

What I'm saying is something like below (only an idea, it even can't
compile), any thoughts?


diff --git a/drivers/irqchip/irq-gic-v3-its.c 
b/drivers/irqchip/irq-gic-v3-its.c
index 2e12bc52e3a2..3450b5e847ca 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -198,7 +198,8 @@ static u16 get_its_list(struct its_vm *vm)
  		if (!is_v4(its))
  			continue;

-		if (vm->vlpi_count[its->list_nr])
+		if (vm->vlpi_count[its->list_nr] ||
+		    gic_requires_eager_mapping())
  			__set_bit(its->list_nr, &its_list);
  	}

@@ -1295,7 +1296,8 @@ static void its_send_vmovp(struct its_vpe *vpe)
  		if (!is_v4(its))
  			continue;

-		if (!vpe->its_vm->vlpi_count[its->list_nr])
+		if (!vpe->its_vm->vlpi_count[its->list_nr] &&
+		    !gic_requires_eager_mapping())
  			continue;

  		desc.its_vmovp_cmd.col = &its->collections[col_id];


Thanks,
Zenghui

> ---
>   drivers/irqchip/irq-gic-v3-its.c | 39 +++++++++++++++++++++++++-------
>   1 file changed, 31 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
> index b65fba67bd85..6277b3e3731f 100644
> --- a/drivers/irqchip/irq-gic-v3-its.c
> +++ b/drivers/irqchip/irq-gic-v3-its.c
> @@ -1586,12 +1586,31 @@ static int its_irq_set_irqchip_state(struct irq_data *d,
>   	return 0;
>   }
>   
> +/*
> + * Two favourable cases:
> + *
> + * (a) Either we have a GICv4.1, and all vPEs have to be mapped at all times
> + *     for vSGI delivery
> + *
> + * (b) Or the ITSs do not use a list map, meaning that VMOVP is cheap enough
> + *     and we're better off mapping all VPEs always
> + *
> + * If neither (a) nor (b) is true, then we map vPEs on demand.
> + *
> + */
> +static bool gic_requires_eager_mapping(void)
> +{
> +	if (!its_list_map || gic_rdists->has_rvpeid)
> +		return true;
> +
> +	return false;
> +}
> +
>   static void its_map_vm(struct its_node *its, struct its_vm *vm)
>   {
>   	unsigned long flags;
>   
> -	/* Not using the ITS list? Everything is always mapped. */
> -	if (!its_list_map)
> +	if (gic_requires_eager_mapping())
>   		return;
>   
>   	raw_spin_lock_irqsave(&vmovp_lock, flags);
> @@ -1625,7 +1644,7 @@ static void its_unmap_vm(struct its_node *its, struct its_vm *vm)
>   	unsigned long flags;
>   
>   	/* Not using the ITS list? Everything is always mapped. */
> -	if (!its_list_map)
> +	if (gic_requires_eager_mapping())
>   		return;
>   
>   	raw_spin_lock_irqsave(&vmovp_lock, flags);
> @@ -4260,8 +4279,12 @@ static int its_vpe_irq_domain_activate(struct irq_domain *domain,
>   	struct its_vpe *vpe = irq_data_get_irq_chip_data(d);
>   	struct its_node *its;
>   
> -	/* If we use the list map, we issue VMAPP on demand... */
> -	if (its_list_map)
> +	/*
> +	 * If we use the list map, we issue VMAPP on demand... Unless
> +	 * we're on a GICv4.1 and we eagerly map the VPE on all ITSs
> +	 * so that VSGIs can work.
> +	 */
> +	if (!gic_requires_eager_mapping())
>   		return 0;
>   
>   	/* Map the VPE to the first possible CPU */
> @@ -4287,10 +4310,10 @@ static void its_vpe_irq_domain_deactivate(struct irq_domain *domain,
>   	struct its_node *its;
>   
>   	/*
> -	 * If we use the list map, we unmap the VPE once no VLPIs are
> -	 * associated with the VM.
> +	 * If we use the list map on GICv4.0, we unmap the VPE once no
> +	 * VLPIs are associated with the VM.
>   	 */
> -	if (its_list_map)
> +	if (!gic_requires_eager_mapping())
>   		return;
>   
>   	list_for_each_entry(its, &its_nodes, entry) {
> 


  parent reply	other threads:[~2020-03-17  2:49 UTC|newest]

Thread overview: 99+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-04 20:33 [PATCH v5 00/23] irqchip/gic-v4: GICv4.1 architecture support Marc Zyngier
2020-03-04 20:33 ` [PATCH v5 01/23] irqchip/gic-v3: Use SGIs without active state if offered Marc Zyngier
2020-03-12  6:30   ` Zenghui Yu
2020-03-12  9:28     ` Marc Zyngier
2020-03-12 12:05       ` Marc Zyngier
2020-03-13  1:39         ` Zenghui Yu
2020-03-12 17:16   ` Auger Eric
2020-03-12 17:23     ` Marc Zyngier
2020-03-04 20:33 ` [PATCH v5 02/23] irqchip/gic-v4.1: Skip absent CPUs while iterating over redistributors Marc Zyngier
2020-03-16 17:10   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 03/23] irqchip/gic-v4.1: Ensure mutual exclusion between vPE affinity change and RD access Marc Zyngier
2020-03-12  6:56   ` Zenghui Yu
2020-03-04 20:33 ` [PATCH v5 04/23] irqchip/gic-v4.1: Wait for completion of redistributor's INVALL operation Marc Zyngier
2020-03-20 14:23   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 05/23] irqchip/gic-v4.1: Ensure mutual exclusion betwen invalidations on the same RD Marc Zyngier
2020-03-12  7:11   ` Zenghui Yu
2020-03-20 14:23   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 06/23] irqchip/gic-v4.1: Advertise support v4.1 to KVM Marc Zyngier
2020-03-16 17:10   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 07/23] irqchip/gic-v4.1: Map the ITS SGIR register page Marc Zyngier
2020-03-16 17:10   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 08/23] irqchip/gic-v4.1: Plumb skeletal VSGI irqchip Marc Zyngier
2020-03-16 17:10   ` Auger Eric
2020-03-19 10:03     ` Marc Zyngier
2020-03-04 20:33 ` [PATCH v5 09/23] irqchip/gic-v4.1: Add initial SGI configuration Marc Zyngier
2020-03-16 17:53   ` Auger Eric
2020-03-17  2:02     ` Zenghui Yu
2020-03-17  8:36       ` Auger Eric
2020-03-19 10:20     ` Marc Zyngier
2020-03-04 20:33 ` [PATCH v5 10/23] irqchip/gic-v4.1: Plumb mask/unmask SGI callbacks Marc Zyngier
2020-03-16 18:15   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 11/23] irqchip/gic-v4.1: Plumb get/set_irqchip_state " Marc Zyngier
2020-03-12  7:41   ` Zenghui Yu
2020-03-16 21:43   ` Auger Eric
2020-03-19 10:27     ` Marc Zyngier
2020-03-04 20:33 ` [PATCH v5 12/23] irqchip/gic-v4.1: Plumb set_vcpu_affinity " Marc Zyngier
2020-03-17 10:35   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 13/23] irqchip/gic-v4.1: Move doorbell management to the GICv4 abstraction layer Marc Zyngier
2020-03-12  8:20   ` Zenghui Yu
2020-03-04 20:33 ` [PATCH v5 14/23] irqchip/gic-v4.1: Add VSGI allocation/teardown Marc Zyngier
2020-03-12  8:06   ` Zenghui Yu
2020-03-04 20:33 ` [PATCH v5 15/23] irqchip/gic-v4.1: Add VSGI property setup Marc Zyngier
2020-03-12  8:09   ` Zenghui Yu
2020-03-17 10:30   ` Auger Eric
2020-03-19 10:57     ` Marc Zyngier
2020-03-04 20:33 ` [PATCH v5 16/23] irqchip/gic-v4.1: Eagerly vmap vPEs Marc Zyngier
2020-03-12  8:12   ` Zenghui Yu
2020-03-17  2:49   ` Zenghui Yu [this message]
2020-03-19 10:55     ` Marc Zyngier
2020-03-20  2:31       ` Zenghui Yu
2020-03-04 20:33 ` [PATCH v5 17/23] KVM: arm64: GICv4.1: Let doorbells be auto-enabled Marc Zyngier
2020-03-12  8:15   ` Zenghui Yu
2020-03-17 11:04   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 18/23] KVM: arm64: GICv4.1: Add direct injection capability to SGI registers Marc Zyngier
2020-03-18  3:28   ` Zenghui Yu
2020-03-20  8:11   ` Auger Eric
2020-03-20 10:05     ` Marc Zyngier
2020-03-20 10:56       ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 19/23] KVM: arm64: GICv4.1: Allow SGIs to switch between HW and SW interrupts Marc Zyngier
2020-03-19 16:16   ` Auger Eric
2020-03-19 19:52     ` Marc Zyngier
2020-03-19 20:13       ` Auger Eric
2020-03-20  9:17         ` Marc Zyngier
2020-03-20  4:22   ` Zenghui Yu
2020-03-04 20:33 ` [PATCH v5 20/23] KVM: arm64: GICv4.1: Plumb SGI implementation selection in the distributor Marc Zyngier
2020-03-18  6:34   ` Zenghui Yu
2020-03-19 12:10     ` Marc Zyngier
2020-03-19 20:38       ` Auger Eric
2020-03-20  3:08         ` Zenghui Yu
2020-03-20  7:59           ` Auger Eric
2020-03-20  9:46             ` Marc Zyngier
2020-03-20 11:09               ` Auger Eric
2020-03-20 11:20                 ` Marc Zyngier
2020-03-20  3:53       ` Zenghui Yu
2020-03-20  9:01         ` Marc Zyngier
2020-03-23  8:11           ` Zenghui Yu
2020-03-23  8:25             ` Marc Zyngier
2020-03-23 12:40               ` Zenghui Yu
2020-03-04 20:33 ` [PATCH v5 21/23] KVM: arm64: GICv4.1: Reload VLPI configuration on distributor enable/disable Marc Zyngier
2020-03-18  3:17   ` Zenghui Yu
2020-03-19 12:18     ` Marc Zyngier
2020-03-04 20:33 ` [PATCH v5 22/23] KVM: arm64: GICv4.1: Allow non-trapping WFI when using HW SGIs Marc Zyngier
2020-03-20  4:23   ` Zenghui Yu
2020-03-20  8:12   ` Auger Eric
2020-03-04 20:33 ` [PATCH v5 23/23] KVM: arm64: GICv4.1: Expose HW-based SGIs in debugfs Marc Zyngier
2020-03-18  3:19   ` Zenghui Yu
2020-03-19 15:05   ` Auger Eric
2020-03-19 15:21     ` Marc Zyngier
2020-03-19 15:43       ` Auger Eric
2020-03-19 16:16         ` Marc Zyngier
2020-03-19 16:17           ` Auger Eric
2020-03-20  4:38       ` Zenghui Yu
2020-03-20  9:09         ` Marc Zyngier
2020-03-20 11:35           ` Zenghui Yu
2020-03-20 11:46             ` Marc Zyngier
2020-03-20 12:09               ` Zenghui Yu
2020-03-05  3:39 ` [PATCH v5 00/23] irqchip/gic-v4: GICv4.1 architecture support Zenghui Yu
2020-03-09  8:17 ` Zenghui Yu
2020-03-09  8:46   ` Marc Zyngier

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=2817cb89-4cc2-549f-6e40-91d941aa8a5f@huawei.com \
    --to=yuzenghui@huawei.com \
    --cc=eric.auger@redhat.com \
    --cc=james.morse@arm.com \
    --cc=jason@lakedaemon.net \
    --cc=julien.thierry.kdev@gmail.com \
    --cc=kvm@vger.kernel.org \
    --cc=kvmarm@lists.cs.columbia.edu \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lorenzo.pieralisi@arm.com \
    --cc=maz@kernel.org \
    --cc=rrichter@marvell.com \
    --cc=suzuki.poulose@arm.com \
    --cc=tglx@linutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).